Research library

DeepLabCut: markerless pose estimation of user-defined body parts with deep learning

Alexander Mathis 1,2, Pranav Mamidanna1, Kevin M. Murthy Mackenzie Weygandt Mathis 1,4,8* and Matthias Bethge1,5,6,7,8 2 , Quantifying behavior is crucial for many applications in neuroscience. Videography provides easy methods for the observation and recording of animal behavior in diverse settings, yet extracting particular aspects of a behavior for further analysis can be highly time consuming. In motor control studies, humans or other animals are often marked with reflective markers to assist with computer-based tracking, but markers are intrusive, and the number and location of the markers must be determined a priori.

Download PDF

Publication details

Authors
Alexander Mathis
Organizations
🇩🇪 Eberhard Karls Universität Tübingen🇺🇸 Harvard University🇺🇸 Columbia University🇩🇪 Max Planck Institute for Biological Cybernetics🇩🇪 Bernstein Center for Computational Neuroscience🇺🇸 Baylor College of Medicine
Year
2018
Type
Journal

Relevancy to Gratheon

This paper is relevant to Gratheon because it informs entrance and behavior analytics in the Gratheon web app, camera-based hive-scanner and computer-vision models. Its methods and findings can be translated into product requirements for reliable field deployments: what should be sensed, how signals should be interpreted, and which uncertainty or validation limits need to be surfaced to beekeepers. For Gratheon, the work is most useful as an evidence-backed design reference for connecting local hive observations with actionable recommendations in the web app while keeping hardware practical for remote apiaries.